• Weaviate Newsletter
  • Posts
  • RAG Tutorials, Multi-Tenancy Course, Python Client v3 API Update & Weaviate Recipes

RAG Tutorials, Multi-Tenancy Course, Python Client v3 API Update & Weaviate Recipes

Author

Femke Plantinga
August 29, 2024

Hello Weaviate Community, 🤗

As summer wraps up, we're excited to bring you fresh learning resources, from a RAG YouTube tutorial series to a new Weaviate Academy course on multi-tenancy. Plus, join us for online and in-person events, like hands-on workshops on chunking strategies for RAG. Heads up: the Python client v3 API is being deprecated soon, so make sure to update your projects!

Let’s dive in. 🚀

Python client v3 API deprecation

All good things have to come to an end.

You probably know that the Weaviate Python client got a glow-up to an extensively re-written (and improved) v4 API around six months ago. Up until now, the v4 client versions had included the older v3 API in order to ease the transition.

Starting in December 2024, the v4 client will no longer include the v3 API (i.e. the weaviate.Client class). This will help us to provide the best developer experience for you, support for the latest features, and clearly separate the two.

The v3 client will continue to get critical security updates and bugfixes for the foreseeable future, but it will not support any new features.

What does this mean for me?

To take advantage of the latest developments on the Weaviate core database, we recommend migrating your codebase to use the v4 API.

Our documentation includes a migration guide here, and many code examples include both v3 and v4 syntax. We will be adding more dedicated resources for you to ease the migration experience.

If you have an existing codebase and Weaviate core database that you expect to remain static, we recommend pinning the version in your requirements file (e.g. requirements.txt), like so: 

  weaviate-client>=3.26.7,<4.0.0

We appreciate that code migration can be cumbersome, but we feel strongly that the end experience and feature set will make your time worthwhile.

If you have specific requests for migration documentation or resources, please reach out through our GitHub repository.

AI data tips, tricks, and tech

New Course: Build RAG apps in Weaviate with Next.js

In just 8 short videos, Daniel will take you through setting up Weaviate in Typescript, connecting to Weaviate cloud, and building a RAG frontend:

How-to

Weaviate Academy

  • 🎓 New Course: Multi-tenancy (MT)

    • Learn how to create high-capacity, lightweight tenants within a Weaviate collection, ideal for SaaS applications where each user’s data is isolated and managed independently. This course covers enabling and configuring multi-tenant collections, as well as handling tenant data, including offloading to cold storage.

In Weaviate Recipes, we share end-to-end notebooks on how to use various Weaviate features and integrations!

This week, we’ve added six new recipes to the repository:

  1. DSPy: Construct a DSPy agent that designs DSPy agents inspired by the  "Automated Design of Agentic Systems" paper.

  2. LangChain: Build a multi-language RAG by multiple PDFs per tenant with LangChain, OpenAI, and Weaviate.

  3. Nvidia: CAGRA demo with Nvidia’s cuVS

  4. Haystack: Learn how to implement query expansion for RAG in Haystack.

  5. Composio: Integrate Composio's Gmail tool with Weaviate to create an agent responding to new messages.

  6. Context Data: Three examples show how to ingest data from Google Cloud Storage, Postgres, and S3 into Weaviate.

AI-Native Development with Guy Podjarny and Bob van Luijt - Weaviate Podcast #102!

Join us to discuss the emerging opportunities in AI-Native Development with Guy Podjarny from Tessl and Weaviate Co-Founder Bob van Luijt!

MIPRO and DSPy with Krista Opsahl-Ong - Weaviate Podcast #103!

Join us to discuss the MIPRO DSPy Optimizer and the state of Automated Prompt Engineering with Krista Opsahl-Ong from Stanford University!

Upcoming events

Deep dive into chunking for RAG with Zain

September 18th | 9 am PDT / 12 pm EDT / 6 pm CEST

RAG applications are rapidly gaining popularity, but they come with new challenges. Common questions include:

  • What chunking strategies exist?

  • Which strategy suits my application?

  • How can I optimize for accuracy, speed, or cost?

In our upcoming workshop, Zain will dive deeper into various chunking strategies, including optimal chunk size, overlap windows, semantic chunking, LLM-based chunking, and leveraging metadata to enhance RAG search results.

AI [in Prod] - Seattle

September 19th

Join us in Seattle to discover how enterprises scale generative AI from prototype to production. Come for the tech talks, stay for the hands-on training.

Confluent Current 2024

Planning to be in Austin on September 17th and 18th? Join our team at Confluent Current (booth #503) to learn about Weaviate and our integration with Confluent Cloud!

We will be delivering two talks in the expo hall, where we’ll showcase a demo app called "Wealingo," which provides real-time, personalized language learning usingWeaviate's vector database and Confluent Cloud's streaming capabilities

Online

In-person events in San Francisco

Remember to check out the provided links for all the details on how to sign up. We can't wait to meet you!

Welcoming new faces

  • Josh joins us from New York 🇺🇸 as an Enterprise Solution Engineer.

  • Spiros is from the UK 🇬🇧 and will be our Security and Compliance Lead.

  • Prajjwal joins us from India 🇮🇳 as a Developer Growth Intern.

We’re hiring!

Check out our open roles:

Have a look at our career page for more roles and opportunities! ✨

Thank you for reading

Have questions about Weaviate, vector databases, documentation, or other topics? Say hi at the Community Slack or join our Weaviate Forum, where you can engage in community conversations. We look forward to your participation over the next two weeks!

Weaviate is open source. Check out our GitHub repository, and don't forget to star us while you're there ⭐

Till the next one,

Femke